Skip to contentSkip to conversation

fstests: add a dio-read-into-mmap and sync race test case

PATCH Superseded Qu Wenruo 3 participants 6 messages Jul 25 – Aug 10 Btrfs file system

v1 of 2 Superseded Latest is v2 →

lore
Share & tools
lore

Qu Wenruo

author
#0 lore raw

PATCH 1/1 fstests: add a dio-read-into-mmap and sync race test case

There is a report that on btrfs, if the following workload are running, btrfs can fail:

- A dio read into a mmaped range
Only the mmap range needs to be on btrfs.
The dio read source makes no difference.

- Sync_range on the mapped range

The btrfs errors include:

- Hang during data writeback
- Filesystem flips RO

The mmap range is dirtied but written back by the sync_range process, then dio read finished and found that the folios are no longer dirty, so dio endio will mark those folios dirty again so that the fs can write them back again.

However for non-experimental btrfs with 4K block size and 4K page size, there is a regression in v7.2 that such case is no longer handled properly, due to the enablement of large folios and removal of cow fixup.
And btrfs can never handle it for bs < ps from day 1.

Add a regression test for it.

---
 .gitignore               |  1 +
 src/Makefile             |  2 +-
 src/dio-read-into-mmap.c | 81 ++++++++++++++++++++++++++++++++++++++++
 tests/generic/799        | 81 ++++++++++++++++++++++++++++++++++++++++
 tests/generic/799.out    |  2 +
 5 files changed, 166 insertions(+), 1 deletion(-)
 create mode 100644 src/dio-read-into-mmap.c
 create mode 100755 tests/generic/799
 create mode 100644 tests/generic/799.out
.gitignore +1 -0
index 0b6b9452..d52ba4ad 100644
@@ -216,6 +216,7 @@ tags
 /src/truncate
 /src/t_btrfs_received_uuid_ioctl
 /src/btrfs_ioctl
+/src/dio-read-into-mmap
 
 # Symlinked files
 /tests/generic/035.out
src/Makefile +1 -1
index 76cf50c3..f64e153a 100644
@@ -21,7 +21,7 @@ TARGETS = dirstress fill fill2 getpagesize holes lstat64 \
 	t_mmap_writev_overlap checkpoint_journal mmap-rw-fault allocstale \
 	t_mmap_cow_memory_failure fake-dump-rootino dio-buf-fault rewinddir-test \
 	readdir-while-renames dio-append-buf-fault dio-write-fsync-same-fd \
-	dio-writeback-race unlink-fsync truncate
+	dio-writeback-race unlink-fsync truncate dio-read-into-mmap
 
 LINUX_TARGETS = xfsctl bstat t_mtab getdevicesize preallo_rw_pattern_reader \
 	preallo_rw_pattern_writer ftrunc trunc fs_perms testx looptest \
src/dio-read-into-mmap.c new file +81 -0

81 lines not on this page · raw message

tests/generic/799 new file +81 -0

81 lines not on this page · raw message

tests/generic/799.out new file +2 -0
new file mode 100644
index 00000000..f3fd9fa2
@@ -0,0 +1,2 @@
+QA output created by 799
+Silence is golden
Signature · 1 line
2.51.2
#1 lore raw
Quoted 25 lines · Qu Wenruo

There is a report that on btrfs, if the following workload are running, btrfs can fail:

- A dio read into a mmaped range
Only the mmap range needs to be on btrfs.
The dio read source makes no difference.

- Sync_range on the mapped range

The btrfs errors include:

- Hang during data writeback
- Filesystem flips RO

The mmap range is dirtied but written back by the sync_range process, then dio read finished and found that the folios are no longer dirty, so dio endio will mark those folios dirty again so that the fs can write them back again.

However for non-experimental btrfs with 4K block size and 4K page size, there is a regression in v7.2 that such case is no longer handled properly, due to the enablement of large folios and removal of cow fixup.
And btrfs can never handle it for bs < ps from day 1.

Quoted 6 lines

Add a regression test for it.

---

Hi Wenruo,

Thanks for this new test!

Quoted 91 lines
 .gitignore               |  1 +
 src/Makefile             |  2 +-
 src/dio-read-into-mmap.c | 81 ++++++++++++++++++++++++++++++++++++++++
 tests/generic/799        | 81 ++++++++++++++++++++++++++++++++++++++++
 tests/generic/799.out    |  2 +
 5 files changed, 166 insertions(+), 1 deletion(-)
 create mode 100644 src/dio-read-into-mmap.c
 create mode 100755 tests/generic/799
 create mode 100644 tests/generic/799.out
.gitignore +1 -0
index 0b6b9452..d52ba4ad 100644
@@ -216,6 +216,7 @@ tags
 /src/truncate
 /src/t_btrfs_received_uuid_ioctl
 /src/btrfs_ioctl
+/src/dio-read-into-mmap
 
 # Symlinked files
 /tests/generic/035.out
src/Makefile +1 -1
index 76cf50c3..f64e153a 100644
@@ -21,7 +21,7 @@ TARGETS = dirstress fill fill2 getpagesize holes lstat64 \
 	t_mmap_writev_overlap checkpoint_journal mmap-rw-fault allocstale \
 	t_mmap_cow_memory_failure fake-dump-rootino dio-buf-fault rewinddir-test \
 	readdir-while-renames dio-append-buf-fault dio-write-fsync-same-fd \
-	dio-writeback-race unlink-fsync truncate
+	dio-writeback-race unlink-fsync truncate dio-read-into-mmap
 
 LINUX_TARGETS = xfsctl bstat t_mtab getdevicesize preallo_rw_pattern_reader \
 	preallo_rw_pattern_writer ftrunc trunc fs_perms testx looptest \
src/dio-read-into-mmap.c new file +50 -0
new file mode 100644
index 00000000..07842c48
@@ -0,0 +1,81 @@
+// SPDX-License-Identifier: GPL-2.0
+// Copyright (c) SUSE S.A.
+
+#include <fcntl.h>
+#include <stdlib.h>
+#include <stdio.h>
+#include <unistd.h>
+#include <string.h>
+#include <errno.h>
+#include <sys/mman.h>
+#include <sys/stat.h>
+#include <err.h>
+
+static int read_source_fd = -1;
+static int mmap_dest_fd = -1;
+static void *buf = NULL;
+static int iosize = 4 * 1024 * 1024;
+
+static void usage()
+{
+	fprintf(stderr,
+	"Usage: dio-read-into-mmap <read_source> <mmap_dest>\n");
+}
+
+int main(int argc, char **argv)
+{
+	int ret = -EINVAL;
+	const int pagesize = sysconf(_SC_PAGESIZE);
+	unsigned int cur = 0;
+
+	if (argc != 3) {
+		usage();
+		goto error;
+	}
+	if (iosize < pagesize) {
+		ret = -EINVAL;
+		fprintf(stderr, "blocksize smaller than pagesize\n");
+		goto error;
+	}
+
+	read_source_fd = open(argv[1], O_RDONLY | O_DIRECT, 0600);
+	if (read_source_fd < 0) {
+		ret = -errno;
+		fprintf(stderr, "failed to open '%s': %m", argv[1]);
+		goto error;
+	}
+	mmap_dest_fd = open(argv[2], O_RDWR, 0600);
+	if (mmap_dest_fd < 0) {
+		ret = -errno;
+		fprintf(stderr, "failed to open '%s': %m", argv[2]);

src/dio-read-into-mmap.c line 60

Quoted 10 lines
src/dio-read-into-mmap.c new file +10 -0
+		goto error;
+	}
+	buf = mmap(NULL, iosize, PROT_WRITE, MAP_SHARED, mmap_dest_fd, 0);
+	if (buf == MAP_FAILED) {
+		buf = NULL;
+		fprintf(stderr, "failed to mmap: %m");
+		return -errno;
+	}
+	while (cur < iosize) {
+		ret = pread(read_source_fd, buf, iosize - cur, cur);

Don't we need "(char *)buf + cur" at here?

Quoted 37 lines
src/dio-read-into-mmap.c new file +21 -0
+		if (ret == 0) {
+			ret = -EINVAL;
+			fprintf(stderr, "reached EOF");
+			goto error;
+		}
+		if (ret < 0) {
+			ret = -errno;
+			fprintf(stderr, "failed to read: %m");
+			goto error;
+		}
+		cur += ret;
+	}
+error:
+	close(read_source_fd);
+	close(mmap_dest_fd);
+	if (buf)
+		munmap(buf, iosize);
+	if (ret < 0)
+		return EXIT_FAILURE;
+	return EXIT_SUCCESS;
+}
tests/generic/799 new file +10 -0
new file mode 100755
index 00000000..b6851f69
@@ -0,0 +1,81 @@
+#! /bin/bash
+# SPDX-License-Identifier: GPL-2.0
+# Copyright (c) 2026 SUSE S.A.  All Rights Reserved.
+#
+# FS QA Test 799
+#
+# Test direct read into a mmaped range meanwhile doing data sync for the mmaped
+# range.
+# Such racy workload should cause direct endio function to mark the folio
+# dirty without going through buffered write nor page_mkwrite().

tests/generic/799 line 20

Quoted 10 lines
tests/generic/799 new file +10 -0
+#
+# Make sure the fs can handle such ->dirty_folio() callback correctly.
+#
+. ./common/preamble
+. ./common/filter
+_begin_fstest auto rw mmap
+
+_require_test_program dio-read-into-mmap
+_require_scratch
+_require_test

_require_odirect ?

tests/generic/799 line 30

Quoted 10 lines
tests/generic/799 new file +10 -0
+
+_fixed_by_fs_commit btrfs xxxxxxxxxxxxxx \
+	"btrfs: trigger cow fixup via dirty_folio()"
+
+_cleanup()
+{
+	cd /
+	[ -n "$mount_pid" ] && kill $mount_pid &> /dev/null
+	[ -n "$remount_pid" ] && kill $remount_pid &> /dev/null
+	wait

rm -r -f $tmp.*

tests/generic/799 line 33

Quoted 3 lines
tests/generic/799 new file +3 -0
+}
+
+trap "_cleanup; exit \$status" SIGINT SIGTERM

If you need extra signal, you can run _register_cleanup, but as you just need INT and TERM, so this line can be removed.

Quoted 17 lines
tests/generic/799 new file +17 -0
+_scratch_mkfs >> $seqres.full
+_scratch_mount
+
+# Create the 4MiB target file on TEST_MNT as the read source.
+$XFS_IO_PROG -f -c "pwrite -i /dev/urandom 0 4M" "$TEST_MNT/dio-read-source" >> $seqres.full
+
+# Another 4MiB target file on SCRATCH_MNT as the mmap dest
+$XFS_IO_PROG -f -c "pwrite 0 4M" "$SCRATCH_MNT/mmap-dest" >> $seqres.full
+
+read_workload()
+{
+	trap "wait; exit" SIGTERM
+	while true; do
+		$here/src/dio-read-into-mmap "$TEST_MNT/dio-read-source" \
+			"$SCRATCH_MNT/mmap-dest" &> /dev/null
+		if [ $? -ne 0 ]; then
+			echo "dio read failed"

tests/generic/799 line 60

Quoted 10 lines
tests/generic/799 new file +10 -0
+			break;
+		fi
+	done
+}
+
+sync_workload()
+{
+	trap "wait; exit" SIGTERM
+	while true; do
+		$XFS_IO_PROG -c "sync_range 0 4m" "$SCRATCH_MNT/mmap-dest"

_require_xfs_io_command sync_range

Quoted 6 lines
tests/generic/799 new file +6 -0
+		if [ $? -ne 0 ]; then
+			echo "sync_range failed"
+			break;
+		fi
+	done
+}

tests/generic/799 line 76

Quoted 10 lines
tests/generic/799 new file +10 -0
+
+read_workload &
+read_pid=$!
+
+sync_workload &
+sync_pid=$!
+
+sleep $((15 * $TIME_FACTOR))
+
+kill "$read_pid" "$sync_pid" &> /dev/null

How about:
kill -TERM -"$read_pid" -"$sync_pid" &> /dev/null
?

As the (man 1 kill) says:

pid
Each pid can be expressed in one of the following ways:
...
-n
where n is larger than 1. All processes in process group n are signaled. When an argument of the form '-n' is given, and it is meant to denote a process group, either a signal must be specified first, or the argument must be preceded by a '--' option, otherwise it will be taken as the signal to send.

I didn't give it a try, but I think this might help to kill the dio-read-into-mmap and sync_range process too.

(same in _cleanup)

tests/generic/799 line 77

Quoted 1 line
tests/generic/799 new file +1 -0
+unset "$read_pid" "$sync_pid"

unset read_pid sync_pid

Quoted 16 lines
tests/generic/799 new file +4 -0
+wait
+
+echo "Silence is golden"
+_exit 0
tests/generic/799.out new file +2 -0
new file mode 100644
index 00000000..f3fd9fa2
@@ -0,0 +1,2 @@
+QA output created by 799
+Silence is golden
Signature · 1 line
2.51.2
#2 lore raw
Quoted 28 lines · Zorro Lang
Quoted 27 lines · Qu Wenruo

There is a report that on btrfs, if the following workload are running, btrfs can fail:

- A dio read into a mmaped range
Only the mmap range needs to be on btrfs.
The dio read source makes no difference.

- Sync_range on the mapped range

The btrfs errors include:

- Hang during data writeback
- Filesystem flips RO

The mmap range is dirtied but written back by the sync_range process, then dio read finished and found that the folios are no longer dirty, so dio endio will mark those folios dirty again so that the fs can write them back again.

However for non-experimental btrfs with 4K block size and 4K page size, there is a regression in v7.2 that such case is no longer handled properly, due to the enablement of large folios and removal of cow fixup.
And btrfs can never handle it for bs < ps from day 1.

Add a regression test for it.

Quoted 8 lines
Quoted 4 lines

---

Hi Wenruo,

Thanks for this new test!

Thanks for the review. Will update the patch to address all the comments and refresh the test number.

But one comment inlined below.

Quoted 19 lines

Review

Quoted 2 lines
(quoted diff) +2 -0
+
+kill "$read_pid" "$sync_pid" &> /dev/null

How about:

   kill -TERM -"$read_pid" -"$sync_pid" &> /dev/null

?

As the (man 1 kill) says:

pid
Each pid can be expressed in one of the following ways:
...
-n
where n is larger than 1. All processes in process group n are signaled. When an argument of the form '-n' is given, and it is meant to denote a process group, either a signal must be specified first, or the argument must be preceded by a '--' option, otherwise it will be taken as the signal to send.

Quoted 2 lines

I didn't give it a try, but I think this might help to kill the dio-read-into-mmap and sync_range process too.

Unfortunately this doesn't seems to work.

If I removed the redirection, it shows the following error:

/home/adam/xfstests/tests/generic/801: line 76: kill: (-1271) - No such process /home/adam/xfstests/tests/generic/801: line 76: kill: (-1272) - No such process

Thus it will not really kill the children processes.

Despite this change, all other comments will be addressed.

Thanks,
Qu

Quoted 24 lines

(same in _cleanup)

Quoted 1 line

+unset "$read_pid" "$sync_pid"

unset read_pid sync_pid

Quoted 16 lines
tests/generic/799 new file +4 -0
+wait
+
+echo "Silence is golden"
+_exit 0
tests/generic/799.out new file +2 -0
new file mode 100644
index 00000000..f3fd9fa2
@@ -0,0 +1,2 @@
+QA output created by 799
+Silence is golden
Signature · 1 line
2.51.2
#3 lore raw
Quoted 71 lines · Qu Wenruo
Quoted 28 lines · Zorro Lang
Quoted 27 lines · Qu Wenruo

There is a report that on btrfs, if the following workload are running, btrfs can fail:

- A dio read into a mmaped range
Only the mmap range needs to be on btrfs.
The dio read source makes no difference.

- Sync_range on the mapped range

The btrfs errors include:

- Hang during data writeback
- Filesystem flips RO

The mmap range is dirtied but written back by the sync_range process, then dio read finished and found that the folios are no longer dirty, so dio endio will mark those folios dirty again so that the fs can write them back again.

However for non-experimental btrfs with 4K block size and 4K page size, there is a regression in v7.2 that such case is no longer handled properly, due to the enablement of large folios and removal of cow fixup.
And btrfs can never handle it for bs < ps from day 1.

Add a regression test for it.

Quoted 8 lines
Quoted 4 lines

---

Hi Wenruo,

Thanks for this new test!

Thanks for the review. Will update the patch to address all the comments and refresh the test number.

But one comment inlined below.

Quoted 19 lines

Review

Quoted 2 lines
(quoted diff) +2 -0
+
+kill "$read_pid" "$sync_pid" &> /dev/null

How about:

   kill -TERM -"$read_pid" -"$sync_pid" &> /dev/null

?

As the (man 1 kill) says:

pid
Each pid can be expressed in one of the following ways:
...
-n
where n is larger than 1. All processes in process group n are signaled. When an argument of the form '-n' is given, and it is meant to denote a process group, either a signal must be specified first, or the argument must be preceded by a '--' option, otherwise it will be taken as the signal to send.

Quoted 2 lines

I didn't give it a try, but I think this might help to kill the dio-read-into-mmap and sync_range process too.

Unfortunately this doesn't seems to work.

If I removed the redirection, it shows the following error:

Quoted 8 lines

/home/adam/xfstests/tests/generic/801: line 76: kill: (-1271) - No such process
/home/adam/xfstests/tests/generic/801: line 76: kill: (-1272) - No such process

Thus it will not really kill the children processes.

Despite this change, all other comments will be addressed.

Hi Wenruo,

Sorry for the confusion! I assumed that approach would work, but from the error message, it seems running "read_workload &" doesn't create a proper Process Group (PGID).

How about this approach instead?

1) kill -TERM "$read_pid" 2>/dev/null

Based on the trap in read_workload(), it should enter "wait" upon receiving SIGTERM, preventing any new iterations of the loop. Then we can run:

2) pkill -P "$read_pid" 2>/dev/null

to clean up any remaining child processes, followed by a wait for the main process to exit.

3) wait "$read_pid"
unset read_pid

(We can wrap this logic in a function and invoke it inside _cleanup)

I hope this can prevent the script from waiting long time, and no orphaned processes are left behind. Feel free to tell me if anyone has better idea :)

Thanks,
Zorro

Quoted 30 lines

Thanks,
Qu

Quoted 24 lines

(same in _cleanup)

Quoted 1 line

+unset "$read_pid" "$sync_pid"

unset read_pid sync_pid

Quoted 16 lines
tests/generic/799 new file +4 -0
+wait
+
+echo "Silence is golden"
+_exit 0
tests/generic/799.out new file +2 -0
new file mode 100644
index 00000000..f3fd9fa2
@@ -0,0 +1,2 @@
+QA output created by 799
+Silence is golden
Signature · 1 line
2.51.2
#4 lore raw
Quoted 91 lines · Zorro Lang
Quoted 71 lines · Qu Wenruo
Quoted 28 lines · Zorro Lang
Quoted 27 lines · Qu Wenruo

There is a report that on btrfs, if the following workload are running, btrfs can fail:

- A dio read into a mmaped range
Only the mmap range needs to be on btrfs.
The dio read source makes no difference.

- Sync_range on the mapped range

The btrfs errors include:

- Hang during data writeback
- Filesystem flips RO

The mmap range is dirtied but written back by the sync_range process, then dio read finished and found that the folios are no longer dirty, so dio endio will mark those folios dirty again so that the fs can write them back again.

However for non-experimental btrfs with 4K block size and 4K page size, there is a regression in v7.2 that such case is no longer handled properly, due to the enablement of large folios and removal of cow fixup.
And btrfs can never handle it for bs < ps from day 1.

Add a regression test for it.

Quoted 8 lines
Quoted 4 lines

---

Hi Wenruo,

Thanks for this new test!

Thanks for the review. Will update the patch to address all the comments and refresh the test number.

But one comment inlined below.

Quoted 19 lines

Review

Quoted 2 lines
(quoted diff) +2 -0
+
+kill "$read_pid" "$sync_pid" &> /dev/null

How about:

    kill -TERM -"$read_pid" -"$sync_pid" &> /dev/null

?

As the (man 1 kill) says:

pid
Each pid can be expressed in one of the following ways:
...
-n
where n is larger than 1. All processes in process group n are signaled. When an argument of the form '-n' is given, and it is meant to denote a process group, either a signal must be specified first, or the argument must be preceded by a '--' option, otherwise it will be taken as the signal to send.

Quoted 2 lines

I didn't give it a try, but I think this might help to kill the dio-read-into-mmap and sync_range process too.

Unfortunately this doesn't seems to work.

If I removed the redirection, it shows the following error:

Quoted 8 lines

/home/adam/xfstests/tests/generic/801: line 76: kill: (-1271) - No such process
/home/adam/xfstests/tests/generic/801: line 76: kill: (-1272) - No such process

Thus it will not really kill the children processes.

Despite this change, all other comments will be addressed.

Hi Wenruo,

Sorry for the confusion! I assumed that approach would work, but from the error message, it seems running "read_workload &" doesn't create a proper Process Group (PGID).

How about this approach instead?

1) kill -TERM "$read_pid" 2>/dev/null

Quoted 7 lines

Based on the trap in read_workload(), it should enter "wait" upon receiving SIGTERM, preventing any new iterations of the loop. Then we can run:

2) pkill -P "$read_pid" 2>/dev/null

to clean up any remaining child processes, followed by a wait for the main process to exit.

Not sure if pkill is really needed, as the last wait will handle it well.

So in short, the existing

kill "$pid"
wait "$pid"
unset pid

is good enough, that's what I did in the v2.

Although the child processes still need to trap the corresponding signals, which is also done in v2.

Finally I have already tested the signal interruption for v2 several times, it all properly catch the signal and terminate the children processes and the main script without any delay.

So I think the one used in v2 is working properly.

Thanks,
Qu

Quoted 43 lines

3) wait "$read_pid"
unset read_pid

(We can wrap this logic in a function and invoke it inside _cleanup)

I hope this can prevent the script from waiting long time, and no orphaned processes are left behind. Feel free to tell me if anyone has better idea :)

Thanks,
Zorro

Quoted 30 lines

Thanks,
Qu

Quoted 24 lines

(same in _cleanup)

Quoted 1 line

+unset "$read_pid" "$sync_pid"

unset read_pid sync_pid

Quoted 16 lines
tests/generic/799 new file +4 -0
+wait
+
+echo "Silence is golden"
+_exit 0
tests/generic/799.out new file +2 -0
new file mode 100644
index 00000000..f3fd9fa2
@@ -0,0 +1,2 @@
+QA output created by 799
+Silence is golden
Signature · 1 line
2.51.2
#5 lore raw
Quoted 107 lines · Qu Wenruo
Quoted 91 lines · Zorro Lang
Quoted 71 lines · Qu Wenruo
Quoted 28 lines · Zorro Lang
Quoted 27 lines · Qu Wenruo

There is a report that on btrfs, if the following workload are running, btrfs can fail:

- A dio read into a mmaped range
Only the mmap range needs to be on btrfs.
The dio read source makes no difference.

- Sync_range on the mapped range

The btrfs errors include:

- Hang during data writeback
- Filesystem flips RO

The mmap range is dirtied but written back by the sync_range process, then dio read finished and found that the folios are no longer dirty, so dio endio will mark those folios dirty again so that the fs can write them back again.

However for non-experimental btrfs with 4K block size and 4K page size, there is a regression in v7.2 that such case is no longer handled properly, due to the enablement of large folios and removal of cow fixup.
And btrfs can never handle it for bs < ps from day 1.

Add a regression test for it.

Quoted 8 lines
Quoted 4 lines

---

Hi Wenruo,

Thanks for this new test!

Thanks for the review. Will update the patch to address all the comments and refresh the test number.

But one comment inlined below.

Quoted 19 lines

Review

Quoted 2 lines
(quoted diff) +2 -0
+
+kill "$read_pid" "$sync_pid" &> /dev/null

How about:

    kill -TERM -"$read_pid" -"$sync_pid" &> /dev/null

?

As the (man 1 kill) says:

pid
Each pid can be expressed in one of the following ways:
...
-n
where n is larger than 1. All processes in process group n are signaled. When an argument of the form '-n' is given, and it is meant to denote a process group, either a signal must be specified first, or the argument must be preceded by a '--' option, otherwise it will be taken as the signal to send.

Quoted 2 lines

I didn't give it a try, but I think this might help to kill the dio-read-into-mmap and sync_range process too.

Unfortunately this doesn't seems to work.

If I removed the redirection, it shows the following error:

Quoted 8 lines

/home/adam/xfstests/tests/generic/801: line 76: kill: (-1271) - No such process
/home/adam/xfstests/tests/generic/801: line 76: kill: (-1272) - No such process

Thus it will not really kill the children processes.

Despite this change, all other comments will be addressed.

Hi Wenruo,

Sorry for the confusion! I assumed that approach would work, but from the error message, it seems running "read_workload &" doesn't create a proper Process Group (PGID).

How about this approach instead?

1) kill -TERM "$read_pid" 2>/dev/null

Quoted 7 lines

Based on the trap in read_workload(), it should enter "wait" upon receiving SIGTERM, preventing any new iterations of the loop. Then we can run:

2) pkill -P "$read_pid" 2>/dev/null

to clean up any remaining child processes, followed by a wait for the main process to exit.

Not sure if pkill is really needed, as the last wait will handle it well.

So in short, the existing

Quoted 8 lines

kill "$pid"
wait "$pid"
unset pid

is good enough, that's what I did in the v2.

Although the child processes still need to trap the corresponding signals, which is also done in v2.

Hi Wenruo,

pkill -P isn't strictly necessary here. The main difference is whether the parent process sends a signal to the child process before calling wait, to help it terminate faster. However, if the child process finishes quickly anyway, sending that signal isn't really required. I'll review v2, thanks.

P.S. xfstests has quite a few places running background bash function like this, And it's handled a bit differently every time. Maybe we should provide a couple of common helpers to handle this kind of execution and cleanup.

Thanks,
Zorro

Quoted 54 lines

Finally I have already tested the signal interruption for v2 several times, it all properly catch the signal and terminate the children processes and the main script without any delay.

So I think the one used in v2 is working properly.

Thanks,
Qu

Quoted 43 lines

3) wait "$read_pid"
unset read_pid

(We can wrap this logic in a function and invoke it inside _cleanup)

I hope this can prevent the script from waiting long time, and no orphaned processes are left behind. Feel free to tell me if anyone has better idea :)

Thanks,
Zorro

Quoted 30 lines

Thanks,
Qu

Quoted 24 lines

(same in _cleanup)

Quoted 1 line

+unset "$read_pid" "$sync_pid"

unset read_pid sync_pid

Quoted 16 lines
tests/generic/799 new file +4 -0
+wait
+
+echo "Silence is golden"
+_exit 0
tests/generic/799.out new file +2 -0
new file mode 100644
index 00000000..f3fd9fa2
@@ -0,0 +1,2 @@
+QA output created by 799
+Silence is golden
Signature · 1 line
2.51.2